卵巢浆液性癌三年生存预后因素的分类树分析

Classification tree analysis in serous ovarian adenocarcinoma patients for prognostic factors associated with three-year survival probability

目的应用分类树模型探讨影响卵巢浆液性癌3年生存的预后因素。方法收集1991年1月—2003年12月北京大学人民医院初次收治、随诊3年并已有临床结局的81例卵巢浆液性癌患者的临床资料，采用分类与回归树（CART）软件建立分类树模型，评价预后因素。结果采用CART软件建立的分类树模型中，年龄是影响卵巢浆液性癌3年生存的最重要的预后因素，其他预后因素包括：手术病理分期、淋巴结转移、术后残余病灶、术后化疗方法及病理分级。替代变量分析显示，手术病理分期和淋巴结转移是年龄的主要替代变量。结论年龄、手术病理分期、淋巴结转移、术后残余病灶、术后化疗方法及病理分级是影响卵巢浆液性癌3年生存预后的主要因素。

Objective To analyze the prognostic factors associated with three-year survival outcome in patients with serous ovarian adenocarcinoma by classification tree. Methods Retrospectively we analyzed 81 cases with serous ovarian adenocarcinoma who had 3-year clinical outcomes and were hospitalized in People＇s Hospital from Jan 1991 to Dec 2003 by classification and regression trees （CART） software. Establish the classification tree. Results Among the factors that were associated with the 3-year survival rate, age was the most important factor, other factors in turn were International Federation of Gynecology and Obstetrics （FIGO） stage, lymphoid metastasis, residual size after operation, chemotherapy and pathologic grade. By substitution variable analysis, it was demonstrated that there was cross interaction between age and residual size as well as age and chemotherapy. Conclusion Age, FIGO stage, lymphoid metastasis, residual size after operation, chemotherapy and pathologic grade are important prognostic factors related with three-year survival probability of serous ovarian adenocarcinoma patients.